TAILIEUCHUNG - Báo cáo khoa học: "Active Learning for Statistical Natural Language Parsing"

It is necessary to have a (large) annotated corpus to build a statistical parser. Acquisition of such a corpus is costly and time-consuming. This paper presents a method to reduce this demand using active learning, which selects what samples to annotate, instead of annotating blindly the whole training corpus. Sample selection for annotation is based upon “representativeness” and “usefulness”. A model-based distance is proposed to measure the difference of two sentences and their most likely parse trees. Based on this distance, the active learning process analyzes the sample distribution by clustering and calculates the density of each sample to. | Active Learning for Statistical Natural Language Parsing Min Tang Spoken Language Systems Group MIT Laboratory for Computer Science Cambridge Massachusetts 02139 USA mtang@ Xiaoqiang Luo Salim Roukos IBM . Watson Research Center Yorktown Heights NY 10598 xiaoluo roukos@ Abstract It is necessary to have a large annotated corpus to build a statistical parser. Acquisition of such a corpus is costly and time-consuming. This paper presents a method to reduce this demand using active learning which selects what samples to annotate instead of annotating blindly the whole training corpus. Sample selection for annotation is based upon representativeness and usefulness . A model-based distance is proposed to measure the difference of two sentences and their most likely parse trees. Based on this distance the active learning process analyzes the sample distribution by clustering and calculates the density of each sample to quantify its representativeness. Further more a sentence is deemed as useful if the existing model is highly uncertain about its parses where uncertainty is measured by various entropy-based scores. Experiments are carried out in the shallow semantic parser of an air travel dialog system. Our result shows that for about the same parsing accuracy we only need to annotate a third of the samples as compared to the usual random selection method. 1 Introduction A prerequisite for building statistical parsers Jelinek et al. 1994 Collins 1996 Ratnaparkhi 1997 Charniak 1997 is the availability of a large corpus of parsed sentences. Acquiring such a corpus is expensive and timeconsuming and is often the bottleneck to build a parser for a new application or domain. The goal of this study is to reduce the amount of annotated sentences and hence the development time required for a statistical parser to achieve a satisfactory performance using active learning. Active learning has been studied in the context of many natural language processing

• Báo cáo khoa học
• Active Learning
• Statistical Natural
• Language Parsing
• báo cáo khoa học
• mô hình ngôn ngữ
• xử lý ngôn ngữ tự nhiên
• Promoting active learning and strategies
• Promoting active learning for students at Ba Ria Vung Tau
• Strategies students at Ba Ria Vung Tau
• Effective learning method
• Current learning styles
• Graduation thesis
• English language graduation thesis
• English class
• Learning in English class
• Student performance in English class
• Journal of technology and science education
• Effects of active learning methodologies
• The students’ emotions
• self efficacy beliefs
• Science distance learning course
• The positive emotions
• Journal of Automation and Control Engineering
• Active learning in motor control and matlab simulation
• Active learning in motor control
• Servo motors of the simulation of neural networks
• Motion of a robot arm
• Enhancing moodle for active learning
• Interactive multimedia
• Interactive media
• Quizventure game
• Moodle native counterpart
• A Combination of Active Learning
• Semi supervised Learning Starting
• Positive and Unlabeled Examples for Word Sense Disambiguation An Empirical Study on Japanese Web Search Query
• báo cáo ngôn ngữ
• Information Sciences
• Intelligent tutoring system
• Supporting active learning
• Predictive parsing learning
• Formal language processing
• BMC Bioinformatics
• Pathological image classification
• Atrous convolution
• Deep learning
• Multiscale feature
• Learning stations
• Active teaching learning
• Inquiry teaching
• Primary education
• Cooperative learning
• Machine learning
• Drug development
• Computational biology
• Drug discovery
• Using Smaller Constituents Rather Than Sentences
• Japanese Dependency Parsing
• ngôn ngữ tự nhiên
• Bucking the Trend
• Large Scale Cost Focused Active Learning
• Statistical Machine Translation
• Phrase based Statistical Language Generation
• Graphical Models
• Good Seed Makes a Good Crop
• Accelerating Active Learning
• Language Modeling
• Active Learning Based Elicitation
• Semi Supervised Word Alignment
• Vamshi Ambati
• An Active Learning Approach
• Finding Related Terms
• David Vickrey
• Multilingual Statistical Machine Translation
• Gholamreza Haffari and Anoop Sarkar
• Evaluating the Impact
• Coder Errors
• Semi Supervised Active Learning
• Sequence Labeling
• Katrin Tomanek and Udo Hahn Jena
• Multi Task Active Learning
• Linguistic Annotations
• Roi Reichart
• Assessing the Costs of Sampling
• Methods in Active Learning for Annotation
• Robbie Haertel
• Active Learning with Confidence
• Mark Dredze
• Koby Crammer
• Domain Adaptation
• Word Sense Disambiguation
• Multi Criteria based Active Learning
• Named Entity Recognition
• Dan Shen